topic 3: high-speed mixing measurements
Post on 16-Oct-2021
1 Views
Preview:
TRANSCRIPT
Topic 3: High-speed mixing measurements
Julien ManinSandia National Laboratories
Artium Technologies
Fifth Workshop of the Engine Combustion Network,
Detroit, Michigan, March 2017
ECN 5: Mixing measurements 2/21March 2017
Motivation and previous work
High-speed Rayleigh scatter laser and imaging systems
Spray vessel and injection system
Rayleigh scatter advanced image processing
Quantitative mixing measurements
Summary and future work
Table of contents
ECN 5: Mixing measurements 3/21March 2017
Motivation and previous work
Mixture preparation is considered the back-bone of combustion
Mixing experiments are difficult under relevant thermodynamic and injection conditions like Spray A
Mixing measurements at Spray A have been performed at Sandiaa few years back
But transient information has been lacking mainly due to equipment limitations
Other issues effectively limited the optical resolution of the system
Despite these limitations, this dataset has been used extensively for model validation beyond the thermal engine application boundaries
New equipment such as high-speed lasers or imaging systems can extend experimental capabilities for mixing measurements under relevant conditions
Transient mixing field is a preliminary step towards predicting ignition
Variance quantification is another important parameter to turbulence modeling
Pickett et al. SAE Int. J. Engines 2011
ECN 5: Mixing measurements 4/21March 2017
Motivation and previous work
High-speed Rayleigh scatter laser and imaging systems
Spray vessel and injection system
Rayleigh scatter advanced image processing
Quantitative mixing measurements
Summary and future work
Table of contents
ECN 5: Mixing measurements 5/21March 2017
High-power pulsed burst laser system
High repetition rate seed laser followed by three multi-pass amplification stages
Q-switched diode-pumped Nd-YAG seed laser operated at 100 kHz
Diode-pumped resonators with 2 and 5 mm diameter rods for the first two stages (Nd-YAG)
Final stage uses 10 mm rods in two legs to reduce beam energy on amplifiers
Over 70 mJ/pulse capabilities (7 kW)
Relay imaging of the beam and spatial filtering done throughout the laser path
KTP crystal used on both legs for harmonic generation (532 nm)
Single leg pulse energy was enough
Relay imaged to sheet-forming optics
20-mm wide laser sheet (approx. 15 to 35 mm from the injector)
Parameter Quantity
Frequency 100 kHz
Burst duration 5 ms
Pulse width 4 – 8 ns
Wavelength 532 nm
Pulse energy 15 mJ
Polarization Horizontal
Laser design team: Scott Bisson, Brian Patterson and Jonathan Frank
ECN 5: Mixing measurements 6/21March 2017
High-speed imaging system
The design and development of a high-speed imaging system was investigated
But CMOS-based high-speed cameras evolve too quickly to justify designing and building a specific imaging system
The selected camera features higher light sensitivity and smaller pixels than the competition
Raw images are corrected for intensitylinearity and lens vignetting for accurate signal quantification prior to post-processing
Parameter Quantity
Framerate 100 kfps
Resolution 384 x 264 pix2
Camera lens 58 mm – f/1.2
Close-up lens 2 diopters
Scale factor 78 µm/pix
Spectral filter 532 nm (BP 10 nm)
Camera mounted perpendicular to the jet to acquire Rayleigh scattered signal through the top-mounted optical access of the vessel
Heat-shielding and liquid cooling system were implemented to keep the lens cooled to avoid thermal stress and image deformation
ECN 5: Mixing measurements 7/21March 2017
Motivation and previous work
High-speed Rayleigh scatter laser and imaging systems
Spray vessel and injection system
Rayleigh scatter advanced image processing
Quantitative mixing measurements
Summary and future work
Table of contents
ECN 5: Mixing measurements 8/21March 2017
Spray vessel and conditions
High-temperature and high-pressure capable optically-accessible combustion vessel
35 MPa – 1800 K peak conditions
Sapphire window for Rayleigh scatter signal acquisition
Fused-silica window slits for laser input and output
Recessed and baffled window design with AR-coating to reduce reflections
Vessel surfaces coated with matte black paint to minimize surface reflections
Spray A inert conditions were tested:
Ambient temperature: 900 K
Ambient pressure: 6.0 MPA
Environment composition: 90 % N2, 6 % CO2, and 4 % H2O, 0 % O2
Spray A injector (serial # 210370) at 150 Mpa, 1.5 ms injection duration
ECN 5: Mixing measurements 9/21March 2017
High-precision fuel system
Rayleigh scatter imaging aims at quantifying molecules, and is dependent to the sixth order on particle size
Liquid droplets or particles in the ambient or fuel will contaminate the signal from the probed molecules
Past experiments suffered from particle contamination from the fuel
Solid particles from nozzle deposit have been observed in the head of the spray
New high-pressure fuel system with high-precision syringe pump
Fuel has been filtered through 100 nm membrane filters to remove particles
High-pressure in-line 0.5 µm filter to stop solid contaminants from the fuel lines
Some particles still remain
Axial distance from injector [mm]
5 10 15 20 25 30 35 40 45 50
-15
-10
-5
0
5
10
15
ECN 5: Mixing measurements 10/21March 2017
Motivation and previous work
High-speed Rayleigh scatter laser and imaging systems
Spray vessel and injection system
Rayleigh scatter advanced image processing
Quantitative mixing measurements
Summary and future work
Table of contents
ECN 5: Mixing measurements 11/21March 2017
Particle removal and signal recovery
The high-precision fuel system considerably reduced the amount of particles and/or scatterers
Image treatment is still necessary to eliminate the remaining particles, which would affect measurement quantities
The image processing routine first identifies the particles present on every image
Two-dimensional local inpainting is then used to recover the image intensity behind each particle
No spatial filtering is applied to the image globally
But inpainting effectively smoothens the region covered by a particle
Most particles being small, the smoothing is spatially limited in practice
This process has demonstrated to leave the rest of the image unaffected
Superior to previously applied method such as spatial filtering and global inpainting
ECN 5: Mixing measurements 12/21March 2017
Dynamic flare compensation
Background flare has a considerable effect on fuel concentration quantification and is laser intensity-dependent
Shot-to-shot laser intensity is important and needs to be accounted for
Flare needs to be adjusted on a spatial and temporal basis for accurate mixing measurements
Flare map has been obtained with vessel empty (no pressure)
A sensing region to monitor flare provided the best results
2-D background flare map is adjusted thanks to the sensingregion
ECN 5: Mixing measurements 13/21March 2017
Spatial laser distribution and beam steering
A first method is based on parallel interpolation and corrects laser intensity variations and beam steering simultaneously
Detects jet boundary and extract the Rayleigh signalintensity from the ambient on both sides of the jet
Parallel interpolation is performed linearly along the laser sheet to produce the ambient reference intensity
A second approach applies a wavelet filter to reducebeam steering effects on signal intensity
Normalizes the signal intensity prior to the jet
Wavelet filtering is applied along the laser path to account for intensity variations in the jet region
Laser intensity varies from shot-to-shot and needs to be quantified
Beam steering also affects intensity
ECN 5: Mixing measurements 14/21March 2017
Mixture and temperature quantification
Rayleigh scattering can be used to quantify small particles (molecules) such as fuel vapor
Proportional to molecule number density and scattering cross-section of a species
The equivalent Rayleigh cross-section of a mixture is proportional to the molar fraction and cross-section of the species
Because the chamber only contains ambient gases and injected fuel: Xf + Xa = 1
The Rayleigh scattered intensity in the ambient is used as reference to self-calibrate the fuel – ambient signal
Adiabatic mixing can be applied to assess the temperature of the fuel and ambient mixtures Tmix in the jet
I = I0 ×W×hopt ×Lopt ×N ×s
smix = X fuel ×s fuel +Xamb ×s amb
Imix
Iamb=
s fuel
s amb
+NambN fuel
NambN fuel
+1
Tamb
Tmix
ECN 5: Mixing measurements 15/21March 2017
Motivation and previous work
High-speed Rayleigh scatter laser and imaging systems
Spray vessel and injection system
Rayleigh scatter advanced image processing
Quantitative mixing measurements
Summary and future work
Table of contents
ECN 5: Mixing measurements 16/21March 2017
Time-resolved instantaneous fields
Time resolved 2-D mixture fraction maps for Spray A
Highlight the turbulent nature of the mixing process in diesel jets
Jet progression shows large structures re-entrainment
Adiabatic temperature maps provides information regarding the temperature of the mixture
Along with mixing parameters, temperature is important to determine autoignition
Relevant for thermodynamic state (transcritical)
ECN 5: Mixing measurements 17/21March 2017
Average mixture and temperature fields
Time-resolved fields averaged over 10 repetitions
Detailed mixing features no longer observed
Mixing distribution seems location-dependent more than transient
Mean mixing fields are useful for comparison to models, especially RANS
Averaged fields previously required weeks of experiments
More repetitions are necessary to reduce statistical variation between injections
ECN 5: Mixing measurements 18/21March 2017
Mixing variability
Time-dependent standard deviation across repetitions shows high variability
Jackknife resampling used to enhance statistical confidence
Combination of repetition repeatability and jet turbulence
Standard deviation over quasi-steady period is lower for a single injection
Variability should be compared to mean mixture
Shows higher relative dispersion on the jet boundaries
ECN 5: Mixing measurements 19/21March 2017
Profile comparisons to previous dataset
Mean mixture fraction axial profile matches that of previous experiments
Good agreement between datasets
Standard error calculated across repetitions
Variations may be expected because of the different injector units used
Radial profiles also show similar mixing quantities
Again, mixture fraction profiles are well within uncertainty for both campaigns
The good match between datasets provides confidence with the high-speed measurements
ECN 5: Mixing measurements 20/21March 2017
Motivation and previous work
High-speed Rayleigh scatter laser and imaging systems
Spray vessel and injection system
Rayleigh scatter advanced image processing
Quantitative mixing measurements
Summary and future work
Table of contents
ECN 5: Mixing measurements 21/21March 2017
Summary and future work
Specific equipment has been assembled to allow planar laser Rayleigh scattering imaging measurements in evaporative diesel jets at high-speed
Advanced image processing methodologies were implemented to mitigate the effects of particle contamination, two-dimensional flare, laser intensity variation and beam steering
Time-resolved mixing and temperature measurements are available for Spray A
Ensemble and temporal average data
Mixing variability
Time-resolved results compare well with previous dataset
Detailed mixing quantities are ready for comparisons to high-fidelity CFD simulations
Scalar dissipation rate
Turbulent length scale
Dissipative structure orientation
Topic 3: High-speed mixing measurements
Julien Maninjmanin@artium.com
Fifth Workshop of the Engine Combustion Network,
Detroit, Michigan, March 2017
Thank you for your attention
top related